Aim 1: Data Feasibility Evaluation over GPC sites.

Objective 1.1: Study cohort extraction and characterization

Inclusion criteria:
  • (IP,IS,EI) visits with length of stay (LOS) >= 2 days;
  • age at visit >= 18 years old
Exclusion criteria:
  • had no documented SCr measurements during admission;
  • had initial SCr greater than or equal to 1.3 mg/dL;
  • developed stage 2 AKI (eGFR <= 15 mL/min per 1.73m^2) initially;
  • pre-existing chronic renal failure (DX);
  • received RRT within 48 hours since admission;
  • burn patients (DRG)
AKI Staging

AKI Stages are defined based on KDIGO:

  • AKI 1: increase in serum creatinine by >=0.3 mg/dL within 48 hours OR 1.5-1.9 times baseline within 7 days;
  • AKI 2: 2.0 to 2.9 times baseline within 7 days;
  • AKI 3: increase in serum creatinine to >= 4.0 mg/dL within 48 hours OR 3.0 times baseline within 7 days
Implementation: Extracting AKI Study Cohort
#extract cohort --Table1
# by default, we assume cdm schema is on the same server as current schema,
cohort<-extract_cohort(conn,
                       remote_CDM=params$remote_CDM,
                       cdm_db_link=config_file$cdm_db_link,
                       cdm_db_name=config_file$cdm_db_name,
                       cdm_db_schema=config_file$cdm_db_schema,
                       start_date="2010-01-01",
                       end_date="2018-12-31",
                       verb=F)

The above codes extracted AKI study cohort based on the “Inclusion” and “Exclusion” criteria specified above. The final output will be automatically saved in the current working directory /d1/home/xsong/AKI_CDM as “Table1.rda”. More details are included in the following consort diagram.

Objective 1.2: Variable Collection and Summaries (Table by Table)

In this section, we will collect variables from PCORNET_CDM tables: DEMOGRAPHIC, ENCOUNTER, VITAL, LAB_RESULT_CM, PRESCRIBING MEDICATION, DIAGNOSIS, PROCEDURE, check data quality and generate variable summaries table by table.

Demographic

Demographic variables include Age (calculated from Birth_Date), Sex, Race, Ethnicity.

Table1 - Demographic Summaries at AKI1, AKI2, AKI3
key value ADMIT AKI1 AKI2 AKI3 NONAKI
Age Group
AGE_GRP 18-25 8061, 6% 816, 4% 105, 5% 75, 5% 7150, 6%
AGE_GRP 26-35 13009, 9% 1316, 7% 161, 7% 129, 8% 11571, 9%
AGE_GRP 36-45 15384, 11% 1765, 9% 257, 11% 200, 13% 13430, 11%
AGE_GRP 46-55 25609, 18% 3206, 16% 408, 18% 277, 18% 22122, 18%
AGE_GRP 56-65 34359, 24% 5058, 25% 615, 27% 427, 27% 28910, 23%
AGE_GRP 66<= 48927, 34% 7677, 39% 760, 33% 453, 29% 40823, 33%
Hispanic
HISPANIC N 136047, 94% 18643, 94% 2170, 94% 1454, 93% 115993, 94%
HISPANIC NI 1070, 1% 140, 1% 20, 1% 19, 1% 917, 1%
HISPANIC Y 8232, 6% 1055, 5% 116, 5% 88, 6% 7096, 6%
Race
RACE 01 538, 0% 77, 0% 15, 1% 7, 0% 450, 0%
RACE 02 1415, 1% 193, 1% 23, 1% 10, 1% 1207, 1%
RACE 03 21020, 14% 3037, 15% 380, 16% 278, 18% 17690, 14%
RACE 04 108, 0% 13, 0% 5, 0% 2, 0% 92, 0%
RACE 05 109233, 75% 14607, 74% 1651, 72% 1091, 70% 93591, 75%
RACE 06 314, 0% 36, 0% 7, 0% 5, 0% 273, 0%
RACE 07 509, 0% 80, 0% 5, 0% 9, 1% 425, 0%
RACE NI 165, 0% 29, 0% 6, 0% 5, 0% 132, 0%
RACE OT 12047, 8% 1766, 9% 214, 9% 154, 10% 10146, 8%
Sex
SEX F 72415, 50% 8564, 43% 1082, 47% 547, 35% 63156, 51%
SEX M 72932, 50% 11274, 57% 1224, 53% 1014, 65% 60848, 49%
SEX NI 2, 0% 0, 0% 0, 0% 0, 0% 2, 0%
Total
TOTAL (%/overall) 145349, 100% 19838, 14% 2306, 2% 1561, 1% 124006, 85%

Demographic characterizations for patients at different AKI stages are summarized in Table 1.

Vital

Vital variables include: Height, Weight, BMI, Blood Pressure (Systolic, Diastolic), Smoking Status.

Table 2a - Vital (HT,WT,BMI,SBP,DBP) Summaries
days_from_admit 1.encounters# 2.records# 3.low_records# 4.high_records# 5a.min 5b.median 5c.mean 5d.sd 5e.cov 5f.max
BMI
0> 30904 40140 0 642 2 28 32 230 7.2 24857
1 137742 137742 0 2983 3 28 32 285 8.9 73000
2 6019 6019 0 122 2 28 30 10 0.3 287
3 4748 4748 0 64 12 28 29 9 0.3 242
4 3756 3756 0 51 13 28 29 8 0.3 174
5 2321 2321 0 36 14 28 30 9 0.3 213
6 2147 2147 0 36 13 28 30 13 0.4 442
7 2336 2336 0 41 14 29 30 8 0.3 105
7< 13348 19426 0 283 4 29 31 140 4.5 19051
overall 140992 218635 0 4258 2 28 32 250 7.8 73000
BP_DIASTOLIC
0> 40800 404763 2754 1006 0 70 71 14 0.2 232
1 115366 752487 8356 7856 0 73 74 17 0.2 235
2 109198 968666 10562 3314 0 69 70 15 0.2 236
3 110050 776417 7080 1741 0 69 70 15 0.2 220
4 95597 639243 5367 1435 1 69 70 15 0.2 206
5 80142 523840 4445 985 0 69 70 14 0.2 200
6 67215 434161 3758 779 5 69 70 14 0.2 209
7 57497 372107 3311 608 0 69 69 14 0.2 216
7< 73764 3518467 38818 5384 0 68 69 14 0.2 237
overall 145308 8390151 84451 23108 0 69 70 15 0.2 237
BP_SYSTOLIC
0> 40800 422493 <11 378 34 124 126 21 0.2 467
1 115366 789136 17 4041 30 127 129 25 0.2 313
2 109198 1026881 16 936 0 123 125 22 0.2 310
3 110050 816349 <11 417 0 123 125 21 0.2 285
4 95597 670250 <11 335 31 124 125 21 0.2 306
5 80142 550025 <11 264 31 124 126 21 0.2 292
6 67215 455833 <11 199 31 124 125 21 0.2 315
7 57497 390751 <11 168 0 124 125 21 0.2 260
7< 73764 3707724 54 1371 0 123 124 20 0.2 467
overall 145308 8829442 119 8109 0 124 125 21 0.2 467
HT
0> 29262 37044 0 <11 2 67 67 4 0.1 97
1 131211 131211 0 <11 1 67 67 5 0.1 113
2 5707 5707 0 0 24 67 67 4 0.1 83
3 4448 4448 0 0 24 67 67 4 0.1 81
4 3569 3569 0 0 34 67 67 4 0.1 81
5 2216 2216 0 0 27 67 67 4 0.1 81
6 2044 2044 0 0 18 67 67 4 0.1 80
7 2227 2227 0 0 48 67 67 4 0.1 81
7< 12976 18722 0 0 6 67 67 4 0.1 83
overall 136713 207188 0 <11 1 67 67 5 0.1 113
WT
0> 31398 40693 0 330 16 179 186 52 0.3 705
1 141659 141659 0 1787 20 180 188 56 0.3 1105
2 6036 6036 0 72 12 180 188 56 0.3 1374
3 4768 4768 0 41 71 179 186 52 0.3 621
4 3765 3765 0 24 75 177 183 50 0.3 567
5 2332 2332 0 17 70 183 189 52 0.3 506
6 2153 2153 0 24 79 184 190 52 0.3 481
7 2341 2341 0 27 75 183 190 54 0.3 668
7< 13373 19456 0 169 19 186 192 52 0.3 620
overall 144642 223203 0 2491 12 180 188 54 0.3 1374

Table 2a identifies extreme values of vitals for height, weight, BMI, and blood pressure, which may suggest systemic errors such as typos, and conversion mistakes.

Table 2b - Vital (SMOKING, TABACCO) Summaries
key_cat 1.patients# 2.encounters# 3.encounters%
SMOKING
SMOKING_01,02 92 94 0%
SMOKING_01,03 538 562 0%
SMOKING_NI 39617 52103 36%
SMOKING_03 23931 37139 26%
SMOKING_08 295 360 0%
SMOKING_02,03 349 357 0%
SMOKING_03,08 33 34 0%
SMOKING_02 21703 33608 23%
SMOKING_03,06 32 33 0%
SMOKING_02,06 25 26 0%
SMOKING_01 14517 20508 14%
SMOKING_06 188 196 0%
SMOKING_01,06 16 16 0%
SMOKING_05 114 143 0%
SMOKING_02,08 13 14 0%
SMOKING_07 87 112 0%
TOBACCO
TOBACCO_NI 53016 71983 50%
TOBACCO_02 41722 66927 46%
TOBACCO_01,03 53 54 0%
TOBACCO_01,02 43 44 0%
TOBACCO_03 2643 4018 3%
TOBACCO_01 1590 2167 1%
TOBACCO_02,03 152 155 0%
TOBACCO_TYPE
TOBACCO_TYPE_02 554 761 1%
TOBACCO_TYPE_NI 53057 72053 50%
TOBACCO_TYPE_01,04 623 644 0%
TOBACCO_TYPE_01 28159 43606 30%
TOBACCO_TYPE_03,04 32 33 0%
TOBACCO_TYPE_02,04 31 31 0%
TOBACCO_TYPE_04 16882 26780 18%
TOBACCO_TYPE_01,03 21 22 0%
TOBACCO_TYPE_02,03 20 21 0%
TOBACCO_TYPE_03 1046 1375 1%

Table 2b identifies unreliable reporting of smoking status. A significant mismatch between smoking and tabacco summaries needs some further investigation.

Labs

A total of 817 LOINC identifiable labs are eligible (NI may present), among which 579 are collected at the day of admission, 773 within 3 days. Figure 1 shows the data density and intensity of labs concepts, which can help identify common labs (e.g. the common labs for this study cohort are 2160-0, NI,17861-6,1963-8,2075-0,…), and labs with very high recording intensity (e.g. NI, [lab_report$key[7]], [lab_report$key[8]]).

Diagnosis

A Total of 280 distinct CCS-grouped diagnoses has been assigned to patients before the encounter of interest. Figure 2 gives an overview of average history of patients’ diagnosis prior to tne encounter of interest as well as the highly frequent historical diagnoses(e.g. 259(Residual codes; unclassified), 257(Other aftercare), 133(Other lower respiratory disease), 95(Other nervous system disorders), 155(Other gastrointestinal disorders), 258(Other screening for suspected conditions (not mental disorders or infectious disease))).

There are 249 distcint CCS-group admission diagnoses for encounters of interest. Figure 3 layouts the ditribution of the admission diagnoses associated with patients’ baseline serum creatinine characteristics (e.g. 259(Abdominal pain), 133(Other lower respiratory disease), 106(Cardiac dysrhythmias) are the most common diagnosis for this study cohort; while patients admitted due to 119(Varicose veins of lower extremity), 157(Acute and unspecified renal failure), 158(Chronic kidney disease ) tends to have lower average baseline SCr.

Procedure

A Total of 15951 distinct total procedures codes have been assigned to patients before the encounter of interest. Figure3 gives an overview of average history of patients’ procedures prior to tne encounter of interest as well as the highly frequent historical procedures they had recieved. It can help identify the common procedures or typical occuring times of precedures (e.g. CH:99285, CH:80053,CH:85025,CH:85025,CH:85025,…). Note that Figure2 and Figure3 may display similar distributions as a result of corrlations between diagnoses and procedures.

note: links for cpt codes may lead to invalid page

Medications

A Total of 11539 distinct RXNORM medication concepts are discovered for the cohort. Figure 5 demonstrates average exposures for drug starting at 1st, 2nd, 3rd,…, 7th and after 7th days since admission. It helps identify typical medciations dispensed (:01) or administered(:02) during the course of stay. (e.g. the typical medications identified are 1807627(150 ML Sodium Chloride 9 MG/ML Injection), 1807627(150 ML Sodium Chloride 9 MG/ML Injection), 1807627(150 ML Sodium Chloride 9 MG/ML Injection); while 1807627(150 ML Sodium Chloride 9 MG/ML Injection), 1807627(150 ML Sodium Chloride 9 MG/ML Injection), 1807627(150 ML Sodium Chloride 9 MG/ML Injection) seems to be exposed for relatively long period on average).